06_integrative_analysis_results.Rmd

---
title: "Integrative analysis of colon and small intestine"
author: "A. Gabor"
date: "9/15/2020"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(tidyverse)
```



```{r load_results_1, include=FALSE}
meta_pkn <- read_csv(file = "./data/meta_prior_knowledge/meta_network_carnival_ready_exch_solved_fullomni_metfiltered.csv")
meta_network_colon <-	read_rds("./data/carnival_inputs/meta_pkn_colon.rds")
meta_network_si <-	read_rds("./data/carnival_inputs/meta_pkn_si.rds")
metabolomics <- read_rds("./data/carnival_inputs/human_metabolomics_data.rds")
metabolomics_raw <- read_rds("./data/5-FU in vitro human/metabolomics/raw_metabolomics_tidy.rds")
transcriptomics <- 
tf_activity <- read_rds("./data/carnival_inputs/human_tf_data.rds")
rna_seq_colon <- read_rds("./data/5-FU in vitro human/transcriptomics/colon/colon_transcriptomics_diffExp_filtered_data.rds")
rna_seq_si <- read_rds("./data/5-FU in vitro human/transcriptomics/SI/SI_transcriptomics_diffExp_filtered_data.rds")
expressed_genes_SI <- read_rds("./data/carnival_inputs/expressed_genes_human_SI.rds")
expressed_genes_colon <- read_rds("./data/carnival_inputs/expressed_genes_human_colon.rds")
```


## Overview of the results

### Prior knowledge 

#### Metabolites

The general prior knowledge contains metabolites, metabolic enzymes, proteins, 
kinases and phosphatases, transcription factors. The resources were collected from 
OmnipathDB, STICHdb and RECON3D. The collection is part of the COSMOS package by
Aurelien Dugourd. 


```{r pressure, echo=FALSE}
pkn_metabolites <- tibble(nodes = c(meta_pkn$source, meta_pkn$target)) %>%
	filter(grepl("XMetab__",x = nodes)) %>%
	mutate(nodes = gsub("__[a-z]__","",x = nodes)) %>% 
	mutate(nodes = gsub("XMetab__","",x = nodes)) %>%
	mutate(nodes = gsub("___","",x = nodes)) %>%
	unique()

pkn_metabolites_pubchem = pkn_metabolites %>% filter(grepl("^[0-9]+$",nodes))

data_matabolites <- unique(metabolomics_raw$`PubChem ID`)

metabolomics_raw %>% filter(`PubChem ID` %in% pkn_metabolites$nodes ) %>%
	filter(`q-value (FDR)` < 0.05) %>% pull(`PubChem ID`) %>% unique() %>% length()

```

The prior knowledge contains `R nrow(pkn_metabolites)` unique metabolites across 
multiple compartments. Only `R nrow(pkn_metabolites_pubchem)` metabolites has a
pubChemID. The data contains `R length(data_matabolites)`  unique 
metabolites with pubchemID and `R sum(data_matabolites %in% pkn_metabolites$nodes)`
are included in the prior knowledge. 


#### Gene expression

For TF activity measurements
```{r}
proteins <- tibble(nodes = c(meta_pkn$source, meta_pkn$target)) %>%
	filter(!grepl("XMetab__",x = nodes)) %>%
	mutate(nodes = gsub("__[a-z]__","",x = nodes)) %>% 
	mutate(nodes = gsub("XMetab__","",x = nodes)) %>%
	mutate(nodes = gsub("___","",x = nodes)) %>%
	unique()
```